You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@mxnet.apache.org by re...@apache.org on 2018/11/23 03:42:17 UTC
[incubator-mxnet] branch master updated: Support full convention in
quantized pooling (#13260)
This is an automated email from the ASF dual-hosted git repository.
reminisce pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git
The following commit(s) were added to refs/heads/master by this push:
new cd0ce3b Support full convention in quantized pooling (#13260)
cd0ce3b is described below
commit cd0ce3b13f416525736316389c55a0b31146220c
Author: Tao Lv <ta...@intel.com>
AuthorDate: Fri Nov 23 11:42:02 2018 +0800
Support full convention in quantized pooling (#13260)
* fix quantized pooling and enable it in INT8 SqueezeNet
* add test
* fix test
* address review comments
* refine the test for quantized pooling
---
example/quantization/imagenet_gen_qsym_mkldnn.py | 6 +-----
src/operator/quantization/quantized_pooling.cc | 23 ++++++++++++++++++-----
tests/python/quantization/test_quantization.py | 18 +++++++++++++-----
3 files changed, 32 insertions(+), 15 deletions(-)
diff --git a/example/quantization/imagenet_gen_qsym_mkldnn.py b/example/quantization/imagenet_gen_qsym_mkldnn.py
index 9056f79..c38019f 100644
--- a/example/quantization/imagenet_gen_qsym_mkldnn.py
+++ b/example/quantization/imagenet_gen_qsym_mkldnn.py
@@ -225,11 +225,7 @@ if __name__ == '__main__':
rgb_mean = '123.68,116.779,103.939'
rgb_std = '58.393, 57.12, 57.375'
calib_layer = lambda name: name.endswith('_output')
- excluded_sym_names += ['squeezenet0_flatten0_flatten0',
- 'squeezenet0_pool0_fwd',
- 'squeezenet0_pool1_fwd',
- 'squeezenet0_pool2_fwd',
- 'squeezenet0_pool3_fwd']
+ excluded_sym_names += ['squeezenet0_flatten0_flatten0']
if exclude_first_conv:
excluded_sym_names += ['squeezenet0_conv0_fwd']
elif args.model == 'mobilenet1.0':
diff --git a/src/operator/quantization/quantized_pooling.cc b/src/operator/quantization/quantized_pooling.cc
index 477830a..8b62db9 100644
--- a/src/operator/quantization/quantized_pooling.cc
+++ b/src/operator/quantization/quantized_pooling.cc
@@ -52,17 +52,30 @@ bool QuantizedPoolingShape(const nnvm::NodeAttrs& attrs,
<< "kernel size (" << param.kernel[1]
<< ") exceeds input (" << dshape[W]
<< " padded to " << (dshape[W] + 2*param.pad[1]) << ")";
- // only support valid convention
+
oshape[N] = dshape[N];
oshape[C] = dshape[C];
if (param.global_pool) {
oshape[H] = 1;
oshape[W] = 1;
} else {
- oshape[H] = 1 + (dshape[H] + 2 * param.pad[0] - param.kernel[0]) /
- param.stride[0];
- oshape[W] = 1 + (dshape[W] + 2 * param.pad[1] - param.kernel[1]) /
- param.stride[1];
+ if (param.pooling_convention == pool_enum::kValid) {
+ oshape[H] = 1 +
+ (dshape[H] + 2 * param.pad[0] - param.kernel[0]) /
+ param.stride[0];
+ oshape[W] = 1 +
+ (dshape[W] + 2 * param.pad[1] - param.kernel[1]) /
+ param.stride[1];
+ } else {
+ oshape[H] = 1 + static_cast<int>(std::ceil(
+ static_cast<float>(dshape[H] + 2 * param.pad[0] -
+ param.kernel[0]) /
+ param.stride[0]));
+ oshape[W] = 1 + static_cast<int>(std::ceil(
+ static_cast<float>(dshape[W] + 2 * param.pad[1] -
+ param.kernel[1]) /
+ param.stride[1]));
+ }
}
SHAPE_ASSIGN_CHECK(*in_shape, 1, TShape{1});
diff --git a/tests/python/quantization/test_quantization.py b/tests/python/quantization/test_quantization.py
index 5ae2c6c..e6212b8 100644
--- a/tests/python/quantization/test_quantization.py
+++ b/tests/python/quantization/test_quantization.py
@@ -214,7 +214,7 @@ def test_quantized_conv():
@with_seed()
def test_quantized_pooling():
- def check_quantized_pooling(data_shape, kernel, pool_type, pad, stride, global_pool, qdtype):
+ def check_quantized_pooling(data_shape, kernel, pool_type, pad, stride, global_pool, qdtype, convention='valid'):
if is_test_for_native_cpu():
print('skipped testing quantized_pooling for native cpu since it is not supported yet')
return
@@ -224,7 +224,8 @@ def test_quantized_pooling():
data = mx.sym.Variable(name='data', shape=data_shape, dtype='float32')
pooling_fp32 = mx.sym.Pooling(data=data, kernel=kernel, pad=pad, stride=stride,
- pool_type=pool_type, global_pool=global_pool, cudnn_off=False)
+ pool_type=pool_type, global_pool=global_pool, cudnn_off=False,
+ pooling_convention=convention)
arg_shapes, _, _ = pooling_fp32.infer_shape(data=data_shape)
arg_names = pooling_fp32.list_arguments()
pooling_fp32_exe = pooling_fp32.simple_bind(ctx=mx.current_context(), grad_req='null')
@@ -242,9 +243,10 @@ def test_quantized_pooling():
min_data = mx.sym.Variable(name='min_data')
max_data = mx.sym.Variable(name='max_data')
quantized_pooling = mx.sym.contrib.quantized_pooling(data=qdata, min_data=min_data,
- max_data=max_data, kernel=kernel,
- pad=pad, stride=stride, pool_type=pool_type,
- global_pool=global_pool)
+ max_data=max_data, kernel=kernel,
+ pad=pad, stride=stride, pool_type=pool_type,
+ global_pool=global_pool,
+ pooling_convention=convention)
pooling_int8_exe = quantized_pooling.simple_bind(ctx=mx.current_context(), grad_req='null')
qarg_names = quantized_pooling.list_arguments()
pooling_int8_exe.arg_dict[qarg_names[0]][:] = pooling_fp32_exe.arg_dict[arg_names[0]].astype(qdtype)
@@ -266,6 +268,12 @@ def test_quantized_pooling():
check_quantized_pooling((3, 512, 7, 7), (7, 7), 'avg', (0, 0), (1, 1), False, qdtype)
check_quantized_pooling((3, 512, 7, 7), (7, 7), 'avg', (0, 0), (1, 1), True, qdtype)
+ check_quantized_pooling((3, 4, 56, 56), (3, 3), 'max', (0, 0), (2, 2), False, qdtype, 'full')
+ check_quantized_pooling((3, 4, 56, 56), (3, 3), 'max', (0, 0), (2, 2), True, qdtype, 'full')
+ check_quantized_pooling((3, 512, 7, 7), (7, 7), 'avg', (0, 0), (1, 1), False, qdtype, 'full')
+ check_quantized_pooling((3, 512, 7, 7), (7, 7), 'avg', (0, 0), (1, 1), True, qdtype, 'full')
+
+
@with_seed()
def test_quantized_fc():
def check_quantized_fc(data_shape, num_hidden, no_bias, qdtype, flatten=True):